Functional Similarity between Archaeal and Bacterial CorA Magnesium Transporters

Abstract

The constitutively expressed CorA Mg²⁺ transporter is the major Mg²⁺ influx system of Salmonella typhimurium and Escherichia coli. Genomic sequence data indicated the presence of a homolog in the archaeal organism Methanococcus jannaschii. The putative M. jannaschii CorA was expressed in an Mg²⁺-transport-deficient strain of S. typhimurium to determine its functional characteristics. The archaeal CorA homolog is a functional Mg²⁺ uptake system when expressed in S. typhimurium and has properties which are highly similar to those of the normal CorA transporter of S. typhimurium despite having a low level of sequence identity with the protein and being expressed in a lipid membrane of quite different composition than normal. This implies that the overall function of the proteins is the same and further suggests that their structures are very similar.

The CorA Mg²⁺ transporter was originally identified phenotypically in Escherichia coli (10, 12) and has been cloned, sequenced, and characterized from Salmonella typhimurium (4, 5, 13). In these organisms, a single corA locus constitutively expresses a polypeptide of 316 amino acids with a large N-terminal periplasmic domain of about 240 amino acids followed by three transmembrane domains at the C terminus. CorA mediates the influx of Mg²⁺ with a K_a of about 20 μM extracellular Mg²⁺. Influx of Ni²⁺ and Co²⁺ is also supported by CorA albeit at extracellular concentrations that are immediately toxic to the cell. At very high extracellular Mg²⁺ concentrations (>1 mM), CorA also mediates Mg²⁺ efflux. No other genetic locus has been identified as being necessary for Mg²⁺ influx via CorA; in contrast, the efflux seen at high extracellular Mg²⁺ concentrations is altered by mutations in the corB, corC, and corD loci and is abolished in a strain carrying mutations at all three loci (2). Two additional Mg²⁺ influx systems in S. typhimurium, MgtA and MgtB, are sibling P-type ATPases with no sequence similarities to CorA. Both are highly regulated by Mg²⁺ concentration in the growth medium and under normal laboratory growth conditions are almost completely repressed by the Mg²⁺ content of the medium (4, 16, 17).

Work by Silver and colleagues has shown that CorA-like Mg²⁺ uptake systems are present in E. coli, Bacillus subtilis, and Rhodobacter capsulatus (6, 11, 12). Using both genomic Southern blot analysis and PCR followed by Southern blot analysis, we showed that corA or a similar gene was present in a wide variety of gram-negative and gram-positive bacteria (14); these results suggest that CorA forms the dominant Mg²⁺ transport system of the domain Bacteria. Recently, several complete microbial genomic sequences have become available including examples within both the Bacteria and the domain Archaea. A CorA homolog has been shown to be present in all of these sequences to date except for those for Mycoplasma genitalium, Mycoplasma pneumoniae, and Borrelia burgdorferi, confirming the ubiquity of this class of Mg²⁺ transporter. However, comparison of these CorA-like proteins suggests that the major sequence conservation is within the second and third transmembrane domains (8, 13). The question therefore arises whether this sequence conservation reflects a functional similarity. We tested this possibility by comparing the Mg²⁺ transport properties of the putative CorA homolog of the archaeon Methanococcus jannaschii (MJ-CorA) and the S. typhimurium CorA (ST-CorA).

The M. jannaschii corA gene complements a growth defect in S. typhimurium.

The genomic sequence of M. jannaschii (1) contains an open reading frame (accession no. L77117) encoding a polypeptide with sequence similarity to the ST-CorA (13). The amino acid sequences share a low identity of about 22% and a similarity of 23% (Fig. 1). Both polypeptides are highly charged, with predicted pIs of 4.59 and 5.02 for ST-CorA and MJ-CorA, respectively. In addition, the alignment shown in Fig. 1 reveals 15 sites within the polypeptides with charge reversals. The large majority of these are changes to positive charges in MJ-CorA, accounting for the difference in pI. Optimal growth conditions for the marine archaeon M. jannaschii are reported to be 85°C at >200 atm of pressure in seawater (7). Of relevance is that the concentration of Mg²⁺ in seawater is generally about 55 mM. Moreover, archaeal membrane lipid composition is markedly different than that in bacteria, with phospholipids containing ether rather than fatty acyl linkages. Thus it was possible that the relatively low level of sequence similarity of these two proteins did not reflect a functional similarity.

FIG. 1.

Sequence alignment of MJ-CorA and ST-CorA. The two polypeptides were aligned by eye by using the following sets of amino acid similarities (with C and P each having no similar amino acids): G, A, S, and T; I, L, V, and M; H, R, and K; D, E, N, and Q; and F, Y, and W. The three C-terminal transmembrane domains of CorA (13) are underlined. Colons indicate identical amino acids, and periods indicate similar amino acids. S. typh., ST-CorA; M. jann., MJ-CorA.

An M. jannaschii genomic DNA clone containing the putative corA gene was obtained from the American Type Culture Collection as construct AMJFH86 carried on pUC18 in E. coli. The plasmid was transferred into S. typhimurium JR501 by transformation for restriction modification and then into strain MM281 (4), yielding strain MM1556. MM281 carries insertions in all three S. typhimurium Mg²⁺ transport systems, has no detectable Mg²⁺ transport, and requires 10 to 100 mM Mg²⁺ for growth. A strain carrying any one of the three S. typhimurium Mg²⁺ transporters requires no supplemental Mg²⁺ in the growth medium. Thus, MM281 is a good screen for expression of a protein capable of mediating Mg²⁺ uptake. The MJ-CorA in MM1556 was compared to the ST-CorA encoded by a plasmid-borne S. typhimurium corA allele (pRS170) in the same MM281 background (MM1278).

Upon transformation into MM281, the M. jannaschii corA gene complemented the Mg²⁺ growth defect of this strain and allowed relatively normal growth on N-minimal medium supplemented with 0.4% glucose and 0.1% Casamino Acids (4, 10) or Luria-Bertani nutrient agar plates or broth. In N-minimal medium without added Mg²⁺, the strain carrying MJ-CorA exhibited a significant lag period before beginning growth but had the same rate of exponential growth as a plasmid-borne ST-CorA (Fig. 2). The strain carrying MJ-CorA also exhibited a slight rightward shift in Mg²⁺ dependence for growth compared to the plasmid-borne ST-CorA (Fig. 2, insert). Thus at the level of growth, the archaeal MJ-CorA is functional as an Mg²⁺ transporter in S. typhimurium.

FIG. 2.

Time and Mg²⁺ dependence of growth. Overnight cultures of MM1278 (containing ST-CorA) and MM1556 (containing MJ-CorA) were grown in supplemented N-minimal medium containing 100 mM Mg²⁺. Cells were washed twice in the same medium without Mg²⁺ before being resuspended in the same medium. Cells were then aliquoted to tubes containing the same medium plus the indicated Mg²⁺ concentration (inset) at an initial optical density at 600 nm (OD₆₀₀) of 0.05. Cell growth was measured as OD₆₀₀. The inset shows the Mg²⁺ dependence of growth measured after 17 h, and the main panel shows the time course of growth at 0.1 mM Mg²⁺. The experiment was repeated, and similar results were obtained.

The M. jannaschii and S. typhimurium CorA transporters have functionally identical properties.

Cation uptake was measured by previously published procedures by using ⁶³Ni²⁺ as an alternative substrate since ²⁸Mg²⁺ is unavailable (3, 15). ST-CorA has an apparent affinity for Ni²⁺ of about 250 μM compared to an affinity for Mg²⁺ of 20 to 30 μM (15). The final Ni²⁺ concentration in the assay buffer for these experiments was 100 μM; because this is well below the apparent affinity for Ni²⁺, measurement of the cation concentration required for half-maximal inhibition of uptake is within a factor of 2 of the actual K_i for that cation.

The apparent affinities of Mg²⁺ for the two systems as estimated by inhibition of the uptake of ⁶³Ni²⁺ (Fig. 3) were virtually identical. Ca²⁺ did not significantly inhibit either CorA system. Co²⁺ is also transported by ST-CorA (5, 12). In contrast to the affinities of Ni²⁺ and Mg²⁺ for the two CorA transporters, Co²⁺ affinity appeared slightly less for MJ-CorA compared to that for ST-CorA (Fig. 4). Mn²⁺ inhibition was difficult to assess accurately because higher concentrations of Mn²⁺ were acutely toxic to whole cells and caused extensive clumping. Inhibition of ⁶³Ni²⁺ uptake by Mn²⁺ in the MJ-CorA transporter system appeared to be similar to that in the ST-CorA system. Overall, these data indicate that the inhibition profile of MJ-CorA is remarkably similar to that of ST-CorA.

FIG. 3.

Mg²⁺ and Ca²⁺ inhibition of transport. Mg²⁺ and Ca²⁺ inhibitions of ⁶³Ni²⁺ uptake were measured as previously described (3, 15). The data shown are normalized to the maximal uptake of each CorA. The data for Mg²⁺ are the averages of three independent experiments while those for Ca²⁺ are from a single experiment. The Ni²⁺ concentration used was 100 μM. Uptake was measured for 5 min with triplicates at each concentration. Uptake by the ST-CorA in the Mg²⁺ experiments was 1.0 nmol of ⁶³Ni²⁺ min⁻¹ unit of optical density at 600 nm (OD₆₀₀)⁻¹; uptake by the MJ-CorA was 0.14 nmol of ⁶³Ni²⁺ min⁻¹ unit of OD₆₀₀⁻¹. Uptake in the Ca²⁺ experiment was comparable to that in the Mg²⁺ experiments.

FIG. 4.

Co²⁺ and Mn²⁺ inhibition of transport. Co²⁺ and Mn²⁺ inhibitions of ⁶³Ni²⁺ uptake were measured as previously described (3, 15). The data shown are normalized to the maximal uptake of each CorA. A single experiment with each cation (with triplicates at each concentration) was performed and was repeated once with similar results. The Ni²⁺ concentration used was 100 μM. Uptake by the ST-CorA was 0.85 nmol of ⁶³Ni²⁺ min⁻¹ unit of optical density at 600 nm (OD₆₀₀)⁻¹; uptake by the MJ-CorA was 0.09 nmol of ⁶³Ni²⁺ min⁻¹ unit of OD₆₀₀⁻¹.

Total uptake by MJ-CorA in MM281 was less than that in MM281 carrying ST-CorA on a plasmid. In various experiments, the total uptake via MJ-CorA was 5 to 15% of that via ST-CorA. Because the Ni²⁺ and Mg²⁺ affinities of the two systems were comparable, this implies that the V_maxs for the two systems were in approximately the same ratio as total uptakes. There are several potential reasons for the apparently lower uptake capacity of the MJ-CorA. First, plasmid copy number for ST-CorA was greater than that for MJ-CorA. Further, although the apparent ribosomal binding site sequence for the archaeal gene is comparable to that of a favorable bacterial sequence, there is no recognizable S. typhimurium promoter adjacent to the M. jannaschii corA allele; transcription was presumably via read-through from a plasmid-borne promoter, likely from the antibiotic promoter. Most importantly, codon usage in the archaeal MJ-CorA is not optimal for S. typhimurium; this might greatly diminish translation compared to that of ST-CorA. For these reasons, we did not attempt to quantitate a maximal transport rate, although taken together, our results suggest that the V_max of MJ-CorA would not be markedly different than that of ST-CorA if expressed on a per polypeptide basis. Thus, the longer lag phase before growth of MM281 carrying MJ-CorA starts is unlikely to be due to an intrinsically poor transport capacity.

Finally, since M. jannaschii is a thermophile with optimal growth at 85°C, the temperature dependence of uptake in S. typhimurium carrying either MJ-CorA or ST-CorA was examined. Although S. typhimurium viability decreases greatly upon exposure to higher temperatures, the cells remain intact and functional for at least 15 min after heating so transport can be measured. Since transport was being measured in isogenic cells differing only in the CorA protein each was expressing, any differences in transport must be due to the stability of the transporter itself in the context of the S. typhimurium membrane. Cells were heated to the indicated temperature for 5 min before addition of ⁶³Ni²⁺ for assessment of transport over the following 5 min. The ST-CorA rapidly lost activity above 44°C while the MJ-CorA was completely stable at 65°C, the highest temperature tested (Fig. 5). This result is reflective of the normal growth temperatures of the two organisms.

FIG. 5.

Temperature stability of ST-CorA and MJ-CorA. The temperature stability of uptake was measured by diluting an aliquot of cells 1:10 into medium prewarmed to the indicated temperatures. After 5 min, Ni²⁺ containing ⁶³Ni²⁺ was added to a final concentration 100 μM. Uptake was measured for 5 min as previously described (3, 15). Uptake at 44°C by the ST-CorA was 1.5 nmol of ⁶³Ni²⁺ min⁻¹ unit of optical density at 600 nm (OD₆₀₀)⁻¹; uptake by the MJ-CorA at 44°C was 0.22 nmol of ⁶³Ni²⁺ min⁻¹ unit of OD₆₀₀⁻¹.

Conclusions.

Three conclusions arise from these results. First, despite the low sequence similarity of the MJ-CorA and ST-CorA, functionally the proteins appear to be essentially identical. Both could mediate sufficient uptake of Mg²⁺ to allow growth of the Mg²⁺-transport-deficient MM281 strain, even when, as is the case with the MJ-CorA carried in S. typhimurium, expression of the protein was not optimal. In addition, their cation inhibition profiles are very similar. This is surprising because as a marine organism M. jannaschii lives in a medium that contains not only high NaCl but >50 mM Mg²⁺. By the results shown in Fig. 3 and 4, uptake of Mg²⁺ by the MJ-CorA should be saturated in its normal environment. In contrast, S. typhimurium grows in a variety of environments, most of which are not rich in Mg²⁺, and thus a high Mg²⁺ affinity might be required. Thus, there is no obvious physiological reason why the MJ-CorA should have such a high affinity, not just for Mg²⁺, but also for other divalent cations.

The ST-CorA can also mediate efflux of Mg²⁺, but only at high extracellular Mg²⁺ concentrations which are saturating for influx (2). Since the MJ-CorA has evolved in an environment with a high concentration of Mg²⁺, it might also mediate efflux rather than influx as a primary function, particularly since it would normally be exposed to very high Mg²⁺ concentrations in seawater. However, if the primary function of the MJ-CorA was as an efflux system, it would likely leak Mg²⁺ extensively from the S. typhimurium cell, thus inhibiting growth markedly. This phenotype is seen with at least one mutant of ST-CorA (16a). In addition, the quite high affinity of MJ-CorA for the various other cations required for influx argues strongly that this protein is not a primary Mg²⁺ efflux system in M. jannaschii. This further suggests that M. jannaschii possesses a separate active efflux mechanism for Mg²⁺.

Second, even with such relatively weak sequence conservation between the two polypeptides, their highly similar functions imply that their overall tertiary structures may be very similar.

Finally, these data confirm the ubiquity of this class of Mg²⁺ transport system. An archaeal CorA homolog is able to function as an Mg²⁺ transporter in a cell type from another kingdom of life. The two other archaeal organisms whose genomes have been completely sequenced, Archeoglobus fulgidus and Pyrococcus horikoshii ot3, both carry a clear CorA homolog according to genome sequence data (8, 9). Our previous data demonstrated the ubiquity of corA-like genes in virtually all gram-positive and gram-negative bacterial species tested (14). Thus, we suggest that a CorA-like Mg²⁺ transporter is the major constitutive Mg²⁺ uptake system of both the Bacteria and the Archaea. It remains to be seen whether similar proteins will be found in eukaryotes.